Hybrid cloud architecture is a subcategory of the multi-cloud approach, where at least one of the interconnected cloud platforms utilised is a private cloud (either physically on-premise, a proprietary virtual cloud or a third party-hosted private cloud facility) and one a public cloud. All cloud environments that form part of a hybrid cloud, or multi-cloud architecture, are connected to allow workloads, data and applications to be seamlessly shared or switched.
A hybrid cloud strategy, then, is the defined organisational goal to both develop a robust hybrid cloud architecture, and migrate workloads onto it. The goal is to leverage the computing power of the public cloud, while maintaining heightened security for business-critical data by keeping it within the confines of a private cloud.
Hybrid cloud strategies are a major trend in cloud adoption. Of the 90% of enterprise-level organisations that have a multi-cloud strategy in 2020, 69% take a hybrid cloud approach. That’s more than two-thirds of enterprises combining public and private clouds. But why is it important to them? What informed their decision-making process? And would your organisation also benefit from a hybrid cloud strategy?
Find out why and for whom the distinction between an all-public multi-cloud strategy and a public-private hybrid strategy is important and hybrid architecture use cases. If you are still evaluating your own needs, you will gain some insight into whether or not a hybrid cloud should or shouldn’t be part of your organisation’s strategy.
Cloud strategies have been mainstream for some time now, even if application migration lags infrastructure development. Gartner says over 1.3 trillion US dollars in IT investments will be directed toward cloud migration by 2022. That’s a lot of lolly, so there must be a good reason. And for companies, cloud migration makes both functional and financial sense. The advantages, for many, are well covered – scalability, agility and the total cost of ownership (TCO) for enterprise software.
Multi and hybrid cloud environments do, however, need to be well managed. We use Kubernetes and Istio to manage applications across clouds (in-house, AWS, Google, Microsoft).
A hybrid cloud is a combination of on-premise, private and public clouds from third-party providers – AWS, Azure etc. The reason why organisations adopt a hybrid cloud strategy is usually quite simple. They want to leverage the advantages of migrating to a cloud architecture but have certain workloads and data that require a higher level of security or proximity.
A hybrid cloud has three principal components:
An alternative hybrid cloud architecture that doesn’t fit the NIST hybrid cloud definition, but we would apply the term to because of the overlapping considerations, is a private, on-premise virtualised environment connected to a public cloud.
Hybrid cloud architecture means a seamless connection between two or more cloud platforms – at least one of which is public and one private. It is that connectivity that makes the hybrid strategy work by allowing for data to be moved from a private to a public cloud and vice versa. Specific workloads and data that are business-critical, eg. intellectual property or user data, and have stricter security requirements, are usually always run on the private cloud.
There are two common approaches to setting up a hybrid cloud architecture.
The traditional method involves linking a public to a private cloud through a VPN, WAN, LAN or APIs. It’s complicated.
The more contemporary approach to creating a hybrid cloud environment doesn’t require APIs etc. to move workloads around. This architecture works through modern IT operating systems implemented across environments. Everything is managed as a unified platform.
When does IT Outsourcing work?
(And when doesn’t it?)
With very few exceptions, digital-first and digitally progressive companies need to have, and implement, a public cloud strategy to remain competitive on price and performance. But not all workloads are compatible with a public cloud infrastructure. Security is the most common reason why a public cloud platform may not be a viable or the best solution. But not the only one.
To better understand the use cases that fit hybrid cloud strategies it is helpful to first look at private cloud use cases. So let’s do both!
The advantages of a public cloud include high availability and price-elastic scalability. That’s great for a majority of workloads, but there are use cases that are best suited to a private cloud. Some of the most common are:
For certain kinds of data, particularly personal data, it can be a legal or regulatory requirement that it not leave a country or region. Other data categories can also come with storage and transfer requirements that are dictated by external regulations or internal governance frameworks.
A private cloud is often the only way an organisation can meet these requirements within a cloud strategy.
Public cloud platforms need to offer services, tools and embody qualities that appeal to a mass market. That’s their business model and how they can offer computing power and storage as a utility. But the flipside to the coin is that this can sometimes mean the specific infrastructure or hardware a given application or workload needs isn’t available on a public cloud. Or is only available on one and an organisation is wary of vendor lock-in.
For example, a specific workload may require a 24GB RAM VM but just two or three CPUs. A public cloud provider may not have an instance type in the middle ground you need and would have to run the workloads on a bigger alternative. If that workload is heavy, running it on a bigger instance type may be less cost-effective than providing your own perfectly fitting alternative within a private cloud.
If low latency is a priority, a local private cloud could speed up access to an application. Network latency can be an issue if users are located at a significant distance from the closest public cloud provider. Or if high-speed access is required for internal users.
Public cloud resources are almost always cost-optimal for elastic and intermittent workloads because you only pay for what you use when you use it. You don’t need to pay to keep your peak capacity requirement available for when it’s needed.
But if you have workloads that are steady in size and are constant, a private cloud could be the more effective solution.
A hybrid cloud strategy is required when an organisation needs or wants to leverage the advantages and neutralise the weaknesses, of both public and private cloud environments.
It’s already been mentioned that regulatory or governance requirements and priorities may necessitate the use of a private cloud for certain workloads and storage of sensitive data. But there could be other components of an app that needs to at least partly run in a private cloud that would be better run on a public cloud.
A well-designed hybrid cloud architecture will mean different workloads within the same application can be split between private and public clouds.
When an organisation launches a new application, especially if it’s taking an agile ‘fail fast and cheaply’ approach, it may not know if, or when, it will be a success. It may be that consistent and stable demand would mean a private cloud is the more cost-effective long-term solution.
But until it has been established that there will be that long-term demand, it may make more sense to use a public cloud environment. This is why a hybrid cloud can be an optimal architecture for organisations that regularly release new applications that may or may not have longevity.
A private cloud that offers the same kind of high availability and geo-redundancy as a public cloud can be very expensive to set up. A geo-redundant architecture would mean a minimum of twice the cost of a single private cloud and data centres located at a distance to the main data centre. Very few organisations can justify that level of cost but a lack of geo-redundancy could leave them badly exposed.
Hybrid cloud architecture can offer a good solution, balancing the priorities of high availability and disaster recovery. If, for example, the production environment is held on a private cloud but recovery environment on a public cloud, hibernating but ready to be spun up if needed. Data can be replicated between both the private and public clouds but other resources only made operational in a disaster scenario.
Source: Full Host
Cloudbursting refers to a secondary cloud(s) picking up workloads when capacity is reached on the primary, usually private, cloud. If demand spikes and the private cloud can’t handle the workload, it ‘cloudbursts’ over to a public cloud, which handles the additional load. But for a seamless transition, the kind of low-latency network connections between private and public clouds that only a well-designed hybrid architecture offers are necessary.
A hybrid cloud strategy means a private can handle the workload most of the time without needing to maintain capacity for occasional spikes. Major public cloud vendors like AWS and Azure offer services specifically targeting this need and which place private clouds next to public cloud platforms for super low latency at the moments the hybrid solution needs to come into play.
These use cases help highlight the end goal of different hybrid cloud strategies. When analysing your own needs, it’s a good idea to start at the final destination of what you need your cloud environment to achieve. And then reverse engineer to arrive at the right strategy.
What do you need? Cost efficiencies, agility, scalability, security, reliability? In what order of priority and to what extent can you, or are ready to, compromise on competing priorities?
K&C - Creating Beautiful Technology Solutions For 20+ Years . Can We Be Your Competitive Edge?
Drop us a line to discuss your needs or next project